The research for this paper was inspired by the realization that even though automation and robotics are being pushed heavily into monocultural industrial farming settings, very little is being done to explore and design automated tools for diversified and agroecological cropping systems. And yet, our experiments with strip and pixel cropping show that effectively upscaling complex and promising agroecological systems in industrialized settings will probably depend on at least some level of automation. Struck by this gap, we started wondering: (how) can agroecological cropping systems be automated?

In this paper (researched and written together with Clemens Driessen from CSPS), we examine the drive towards automation in monocultural industrial agriculture and question what automation in diversified agroecological cropping systems might look like, using pixel cropping as a case study. Around the FSE pixel cropping field trials, we brought together a range of actors (researchers, engineers, designers, farmers, students, and more) in a series of interactive happenings where we explored the agronomic, ecological, technical, and socio-political challenges of designing and employing a pixel cropping robot. We learned from these happenings that there exists wide a spectrum of imaginations for how automated tools might be appropriately used in agroecological settings, ranging from fully automated visions, to collaborative scenarios, to not at all. In our discussion we analyze this range of outcomes and consider what it might mean (technically, socially) to imbue the ethos of agroecology into automated farming tools, the possibilities for hybrid forms of engagement with automation, and the implications for tool design processes. (open acces here)